👋 Hi! I'm Bibin Wilson. In each TL;DR edition, I share practical DevOps, Kubernetes, Cloud, and MLOps resources to help you stay up to date. Subscribe here to get every TL;DR edition in your inbox.

This Week in DevOpsCube ☕

  1. Learn Kubeflow Pipelines Architecture in 9 Mins.

  2. Set up Kubeflow Pipelines on Kubernetes(step-by-step guide)

  3. How Airbnb Built a Dynamic Configuration Sidecar for Kubernetes.

  4. How Netflix Built Its In-House LLM Serving Platform

  5. Istio Architecture Explained

  6. DevOps Remote Jobs

and more..

☕ Grab a coffee and catch up on this week’s DevOps, MLOps, and AI insights and resources.

📢 Latest in DevOps

  • AWS Introduces CloudFormation Express Mode - AWS introduced CloudFormation Express Mode, enabling infrastructure deployments to complete up to 4× faster by finishing once resource configuration is applied instead of waiting for full resource stabilization.

  • HashiCorp Introduces tfpolicy for Terraform - HashiCorp introduces tfpolicy, a new HCL-based policy framework that brings infrastructure governance and policy enforcement directly into Terraform workflows.

  • Linux Foundation Launches Akrites - The Linux Foundation launched Akrites, which helps organizations find, fix, and address vulnerabilities in critical open source software before AI-powered attacks can exploit them.

  • AWS Introduces Loom for AI Agent Platforms - AWS introduces Loom, an open-source platform for securely building, deploying, and managing AI agents with enterprise-grade governance, identity, and lifecycle management.

🎯 Learn Kubeflow Pipelines Architecture in 9 Minutes

In this guide, I have covered:

- Kubeflow stack overview
- Learn how Kubeflow Pipelines work
- Role of Argo Workflows in KFP execution
- Complete KFP architecture and execution flow
- How KFP creates and manages task pods

and more..

👉 Read it Here: Kubeflow Pipelines Architecture

📖 Worth Reading This Week

⚙️ Setup Kubeflow Pipelines on Kubernetes

Kubeflow is no longer just an MLOps platform.

It is now a cloud-native AI platform to build and manage GenAI, LLMOps, distributed AI training, LLM fine-tuning etc..

So, if your goal is to move into AI infrastructure, MLOps, or LLMOps, we think Kubeflow is one of the useful platforms you can learn today.

So how to get started?

The best way is to set it up and run one yourself.

We have a hands-on guide that covers,

- Kubeflow Pipelines setup on Kubernetes
- Creating pipeline using the Kubeflow SDK
- Triggering and monitor a pipeline run

With Kubernetes ImageVolume, you can treat data like container images.

👉 Read it Here: Setup Kubeflow Pipelines on Kubernetes

🤖 AI Infrastructure

  • How Yelp did ML Model Training - A case study on Yelp's Training Orchestrator, a unified platform that standardizes ML model training, scheduling, resource management, and experiment execution across teams.

  • How Netflix Built Its In-House LLM Serving Platform - A deep dive into Netflix's in-house LLM serving platform, covering model deployment, inference, API design, deployment strategies, and production-scale serving architecture.

  • How Anthropic Secures AI Agents with Containment - Anthropic explains how it secures AI agents using containers, sandboxes, virtual machines, and network controls to limit what agents can access and reduce the impact of failures or attacks.

  • Using AI to Investigate Service Outages Faster - Expedia shares how it built an AI-powered system that analyzes observability data to help engineers investigate root causes and troubleshoot production issues faster.

  • Measuring AI Model Serving the Right Way - This article explains why AI applications should focus on serving more user requests within the expected response time instead of simply processing the highest number of requests.

🤖 Istio Architecture Explained

By the end of this guide, you will learn:

- Why Istio needed a new architecture
- Overview of Istio architecture
- Deep dive into key components: Istiod, Ztunnel, Waypoint Proxy, and CNI
- Is Ztunnel a single point of failure?
- Business use cases of Ambient Mesh, including cost benefits.

👉 Read it Here: Istio Architecture Explained

🛠️ DevOps Tool of the Week (Hami)

It's an open-source Kubernetes middleware that lets you share, isolate, and schedule GPUs and other AI accelerators across pods without touching your application code.

Here's what it does👇

- Splits one GPU into smaller pieces by memory, core count, or number of devices
- Enforces hard memory and compute limits per workload, so one pod can't eat into another's share
- Places jobs on GPUs smartly, either packing them together to save space or spreading them out to avoid slowdowns
- Manages NVIDIA GPUs alongside NPUs, DCUs, MLUs, and other accelerator types through one scheduling workflow
- Runs multi-tenant training, inference, or notebook workloads on the same shared GPU pool.

👉 GitHub Repo: Hami

🛠️ Scaling Amazon EKS to 100,000 Nodes

As AI workloads continue to grow, Kubernetes clusters are being pushed to unprecedented scale. AWS shares the engineering innovations behind Amazon EKS that enable clusters to scale to 100,000 nodes, supporting some of the world's largest AI and HPC workloads.

The engineering highlights include:

  • A re-architected etcd with AWS's internal journal system to improve scalability and API performance.

  • In-memory storage and key-space partitioning to significantly increase Kubernetes API throughput.

  • Network optimizations such as prefix-based IP allocation to efficiently support massive pod deployments.

  • Control plane and scheduler enhancements to reliably manage hundreds of thousands of nodes.

  • Karpenter and infrastructure improvements designed to handle large-scale AI training and inference workloads.

This is a must-read for anyone interested in large-scale Kubernetes architecture and the engineering behind running AI infrastructure at cloud scale.

👉 Read it Here: Under the Hood: Amazon EKS Ultra-Scale Clusters

🎓 Complete Kubernetes & CKA Course

10,000+ engineers have learned through DevOpsCube courses.

From container fundamentals to CKA preparation, every course is self-paced and grounded in real-world scenarios. This is not a long video lecture series.

The CKA course is text-based, illustration-rich, and designed for faster learning and quick revision whenever you need it.

GET 50% OFF CKA COURSE

👉 Coupon: Use code COMMUNITY50 at checkout.

❝

Note: This is a community-only offer! Make use of it!

💼 DevOps Remote Jobs

  1. Solvios Technology - DevOps / AWS Systems Administrator (3-5+ Yrs)

  2. CloudiQS - AI DevOps Engineer (4-8+ Yrs)

  3. MindBrain - AWS Cloud with DevOps Engineer

  4. Code & Theory - Senior Engineer, MLOps (6-10 Yrs)

  5. Capgemini - MLOps Engineer (5-10 Yrs)

  6. OVI - AIOps/MLOps Engineer(4+ Yrs)

  7. CloudSufi - Lead - Software Engineer MLOps

  8. Anaqua - Senior Platform Engineer

  9. Porch - Senior AIOps Engineer(4+ Yrs)

  10. Nutanix - SRE II (4-6 Yrs)

Subscribe to Learn DevOps & MLOps

20,000+ trusted Engineers read our newsletter, where we simplify DevOps, MLOps and Kubernetes into practical, easy-to-follow guides.

Or subscribe here

DevOpsCube team:

Lead Editor: Bibin Wilson / Designer: Aljin / Tech Researcher: Aswin

Reply

Avatar

or to participate

Keep Reading

© 2026 DevOpsCube.
beehiivPowered by beehiiv